MOD: A Deep Mixture Model with Online Knowledge Distillation for Large Scale Video Temporal Concept Localization

Abstract

In this paper, we present and discuss a deep mixture model with online knowledge distillation (MOD) for large-scale video temporal concept localization, which is ranked 3rd in the 3rd YouTube-8M Video Understanding Challenge. Specifically, we find that by enabling knowledge sharing with online distillation, fintuning a mixture model on a smaller dataset can achieve better evaluation performance. Based on this observation, in our final solution, we trained and fintuned 12 NeXtVLAD models in parallel with a 2-layer online distillation structure. The experimental results show that the proposed distillation structure can effectively avoid overfitting and shows superior generalization performance. The code is publicly available at: https://github.com/linrongc/solution_youtube8m_v3

arXiv commentICCV 2019 YouTube8M workshop

References in corpus (8)

• Distilling the Knowledge in a Neural Network
• Two-Stream Convolutional Networks for Action Recognition in Videos
• Convolutional Two-Stream Network Fusion for Video Action Recognition
• Large scale distributed neural network training through online distillation
• Deep Mutual Learning
• Fusing Multi-Stream Deep Networks for Video Classification
• Non-local NetVLAD Encoding for Video Classification
• NeXtVLAD: An Efficient Neural Network to Aggregate Frame-level Features for Large-scale Video Classification